Foreign Relations of the United States, 1945–1950, Emergence of the Intelligence Establishment

369. Memorandum From the Assistant Director for Reports and Estimates (Babbitt) to Director of Central Intelligence Hillenkoetter⁰

Washington, February 17, 1949.

SUBJECT

• “Coordination of Intelligence”

REFERENCE

• Memorandum, 8 February 1949, to Director from Chief, ICAPS¹

1. Reference memorandum lends considerable weight to the recommendations concerning the functions of ICAPS contained in ORE’s comments on the Dulles Report.² It should be pointed out that ICAPS, in reference memorandum, is representing the point of view of OIR and is passing on this point of view to the Director without consultation with the interested Office in CIA, to-wit: ORE.

2. Reference memorandum refers, first, to intelligence memoranda, and second, to the monthly Review of the World Situation. In regard to the IM’s, as the Director is well aware, requests for these reports come in from a variety of sources but in almost all cases it can be said that the normal procedure is “Urgent” because of the short time factor permitted. ORE is currently informing OIR of the subject of each IM as expeditiously as possible. Such notices of the subject matter and scope of reference cannot be done by telephone for security reasons in a vast majority of cases. Delays in communication being what they are, it is quite true that OIR frequently gets its information after the report is prepared and sometimes even after it has been forwarded to the Director. Attempts are being made, whenever possible, to give advance information to OIR as well as to the other IAC agencies concerned.

3. Paragraph 5 of reference memorandum more than implies that ORE is not “doing its part in accordance with the established regulations.” In spite of the fact that the regulations prescribe a cumbersome and time-consuming procedure, ORE is endeavoring to comply with them as fully as possible. It is obvious that the observance of the “normal” coordination process, or even the “urgent” one, would in many cases place the Director in the position of having to reply to a request from an officer of Cabinet rank that while CIA had the information requested, it was unable to forward it until the departments had been consulted.

4. The reference, paragraph 6, implicitly recommends that the CIA Series³ be subjected to the same processes of coordination as are required in the case of the ORE Series.⁴ The basis of this recommendation is indicated in paragraph 7 which says:

… the basic law and regulation under which we function give to CIA the responsibility for only national intelligence, and the method of setting up national intelligence is participation by the various established intelligence agencies.…

CIA is also specifically authorized by “regulation” to produce current intelligence in NSCID #3, paragraph 1b(2).

5. The concept of national intelligence indicated in the foregoing quotation is obviously taken from the Dulles Report, where, I believe, it finds official expression for the first time. It is entitled to respectful consideration, but is, in my opinion, the most controversial aspect of the Dulles Report and certainly has not yet the force of “basic law and regulation.”

6. This controversial concept equates “national intelligence” to “coordinated intelligence” in terms that make “coordinated intelligence” mean nothing more than “joint intelligence.” In the minds of those who took part in the early post-war discussions and conferences on the subject, the central intelligence organization to be set up was to produce, as “strategic and national policy intelligence,” something above and beyond joint intelligence. “Strategic and national policy intelligence” (now called “national intelligence” for short) was understood in terms of the function such intelligence was intended to serve, not in terms of a particular method of production. The essential idea with respect to the production of such intelligence was expressed in the term “final synthesis”: it was to be an authoritative final evaluation and synthesis of all available intelligence, free from the influence of departmental bias. A process of “coordination” was retained, not because joint intelligence was the end in view, but as a means of discovering and noting any departmental position substantially different from the essentially independent final analysis and interpretation of CIA.

7. This clear concept has been somewhat lost sight of since June 1946. It is still valid, however, and is consonant with the existing “basic law and regulation.” Moreover, the Dulles–ICAPS concept finds no explicit support in the existing law and regulations.

8. The law says only that CIA shall “correlate and evaluate intelligence relating to the national security.” It makes no stipulation with respect to “coordination”: it is CIA that is to correlate and evaluate, language originally used to express the concept of independent final analysis and interpretation set forth above.

9. NSCID No. 1 defines “national intelligence” in terms of the national security (its function), not in terms of “coordination” or any other process of production. It specifies that, insofar as is practicable, CIA shall not duplicate departmental research, but shall make use of departmental facilities and production. It says nothing whatever about any coordination of interpretation.

10. NSCID No. 3 describes “national intelligence” as “integrated” (not “coordinated”) departmental intelligence. This distinction appears to be deliberate and points toward the original concept (“final synthesis”) rather than the Dulles-ICAPS concept of CIA’s function. Coordination is referred to only with respect to the procurement of departmental contributions for CIA consideration, not with respect to any process of joint interpretation.

11. DCI 3/1, intended to “facilitate departmental participation in the preparation of national intelligence,” is the highest law or regulation pertinent to the subject and the reference. In view of the preceding demonstration, its relevant provisions must be regarded as a matter of current policy rather than of compliance with the law or with NSC direction. DCI 3/1 prescribes various procedures for obtaining departmental contributions to and concurrence or dissent on “national intelligence reports and estimates” (the SR⁵ and ORE Series). It specifically provides that coordination is not required with respect to current intelligence.

12. DCI 3/1, dated 8 July 1949, was contemporary with the tenth number in the CIA Series (CIA 7–48). It was specifically understood at that time that the procedures prescribed therein with respect to the ORE Series were not intended to apply to the CIA Series, which, by mutual agreement with the departmental agencies, was already accepted, for these purposes, as current intelligence.

13. This agreement was not reached on any basis of abstract consideration, but as a result of practical experience. Initially we undertook to coordinate items in the CIA Series as though they were items in the ORE Series. CIA-1 (September 1947) was so coordinated, at a cost in time and effort, to the agencies as well as to ourselves, far in excess of any resultant benefit. CIA-2 was not coordinated, for absolute lack of any time in which to do so. I do not recall precisely when the agencies begged off. They may well have been prompted to do so by their relief on those occasions when coordination proved impossible to accomplish for want of time. In any case, I do recall distinctly that agreement to regard the CIA Series as current intelligence was reached on agency initiative and was as much for the relief of agency analysts as for our benefit. If any agency representative now wants to resume coordination, he is presumably arguing in the abstract without appreciation of the practical consequences of his proposal.

14. This observation has particular application in the case of the Department of State. One obstacle to the efficient coordination of any appreciation of global scope, such as items in the CIA Series, is the lack of any unit in the IAC agencies comparable to the Global Survey Group, ORE. To coordinate with us in such a case OIR has to employ a team of half a dozen area specialists. Each such specialist is inclined to demand that his area be treated as though it were the sole subject of consideration. The net result is vexatious delay and diversion of effort for ORE and a much greater loss in man-hours for OIR, without substantial effect upon the tenor of the estimate.

15. The Series is presently prepared with cognizance of departmental views as conveyed to us through various media and in discussions relative to estimates in the ORE Series. Specific prior consultation with departmental specialists would consume their time and ours without contributing materially to the preparation of the initial draft. Moreover, experience shows that the comment elicited in the process of coordinating a draft of such scope and character as an item in the CIA Series is essentially editorial in character and without significant substantive effect.

16. It is, of course, quite feasible to coordinate items in the CIA Series as though they were items in the ORE Series, but it can be done only at considerable cost in loss of timeliness and in terms of man-hours expended, without commensurate gain. Under present procedures, which include elaborate coordination within ORE, it is necessary to write in terms of the situation existing ten days in advance of the publication date. The imposition of external coordination would require allowance of at least an additional week, very definitely impairing the timeliness of the appreciation as of its date of publication. (In this connection, ICAPS should be advised that there is not, and cannot be, in real life any such thing as a meaningful oral coordination.) Moreover, in view of the fact that the Global Survey Group is already experiencing difficulty in giving proper attention to its many and various commitments, the additional consumption of time (approximately one man-week) would require the provision of an additional member to the Group. If consultation and coordination with the departmental agencies (including consultation with half a dozen area specialists in OIR) is the controlling consideration, the best procedure would be to set up a permanent interdepartmental committee to produce the monthly review for the Security Council.

17. The issue is, in essence, whether the CIA Series is intended or desired to be (as we suppose) a responsible synthesis and interpretation of the developing global situation, written with cognizance of departmental views, but with independent judgment, or merely a routine joint intelligence periodical.

18. It is recommended that no action be taken on reference memorandum pending a basic policy decision by the National Security Council with respect to the Dulles Report.

Theodore Babbitt ⁶

0. Source: Central Intelligence Agency Historical Files, HS/HC–100, Item 17. Confidential. The source text is a transcript made for the CIA Historian on October 1, 1953. This memorandum draws heavily on a February 16 memorandum from Ludwell L. Montague to Babbitt. (Ibid., HS/HC–450, Item 30)See the Supplement.
1. Document 363.
2. Document 366.
3. Reference is to a monthly CIA publication entitled Review of the World Situation as it Relates to the Security of the United States, which was regularly considered at NSC meetings during 1947–1950. The first issue, which carried the designation “CIA 1,” was prepared for the NSC’s first meeting on September 26, 1947. See the Supplement for selected issues.
4. Reference is to intelligence estimates and coordinated intelligence reports, which at that time were identified by a series designator that included the prefix “ORE,” e.g., ORE 3–48, ORE 1–49, etc.
5. Situation Reports, a series of periodic reports on individual countries.
6. Printed from a copy that bears this typed signature.